Unity Catalog: one governance layer under Databricks data and AI assets

Unity Catalog sits under data and AI interactions in a Databricks workspace: enforcing access when querying tables or calling models, tracking lineage, and logging for audit. Objects are managed through Catalog Explorer, SQL, the CLI, and REST APIs. Workspaces created after November 8, 2023 get it automatically; older workspaces can upgrade. Every governed asset is a securable object with permissions for users, service principals, or groups. Tables, views, volumes, functions, models, and services use a three-level namespace (catalog.schema.object). Tables and volumes are managed (governance plus storage lifecycle) or external (governance only). Other objects such as storage credentials, external locations, connections, and shares sit under the metastore. Capabilities include privileges, attribute-based policies, row and column filters, and discovery via Catalog Explorer. An open-source Unity Catalog exists on GitHub.

Platform and security teams need one permission and lineage model across tables, files, and ML assets instead of separate tool-by-tool controls.

The securable object hierarchy is shared structure and policy—namespaced meaning for data and AI assets that every query and model call must respect.